There is evidence from adoptive T cell transfer, high dose interferon and anti-CTLA-4 trials that immune system can control malignant melanoma. The question arises whether we can elicit long-lived protective CD8+ T cell immunity that will prevent relapse in patients with high risk disease through vaccination with melanoma antigen-loaded dendritic cells (DCs) without generating autoimmune complications by focusing the immune response on melanoma antigens. Our results from clinical trials testing first generation dendritic cell (DC) vaccines in patients with stage IV melanoma provide a proof-of-principle that therapeutic long-lived melanoma antigen-specific CD8+ T cell immunity can be elicited by ex vivo generated DC vaccines. Distinct DC subsets exist that generate distinct types of immune responses. This concept is fundamental for the development of effective cancer vaccines. We found that IL15- DCs are more efficient than IL4-DCs in priming and maturation of melanoma antigen-specific CTLs. We propose to test the immunogenicity of IL15-DCs in patients with high risk resected stage IIIc/stage IV melanoma. These patients are likely not to have immunosuppression driven by bulky tumors, which represents a significant impediment to induction of tumor specific immune responses. Our hypothesis is that IL15-DCs will generate long-lived melanoma-antigen specific memory T cells in patients with high risk melanoma. Aim 1 is a phase I/II single-arm clinical trial in which patients will receive IL15-DCs loaded with melanoma antigens and with control antigens. Aim 2 will determine the breadth of CD8+ T cell responses. Aim 3 will determine whether IL15-DC vaccine generates high-affinity long-lived melanoma-specific CD8+ T cells able to reject tumors. If we conclude on strong immunogenicity with acceptable toxicity, in future studies IL15-DCs will be tested in randomized trials with clinical efficacy endpoints in i) high-risk patients to measure disease-free survival; and ii) in patients with measurable disease to determine the rate of tumor regression.